Nonfinancial Performance Measures, Externalities and Target Setting A Comparative Case Study of Resolutions Through Planning Hansen2010

Management Accounting Research 21 (2010) 1739
Contents lists available at ScienceDirect
Management Accounting Research

journal homepage: www.elsevier.com/locate/mar
Nonnancial performance measures, externalities and target setting:

A comparative case study of resolutions through planning
Allan Hansen
Copenhagen Business School, Department of Operations Management, Solbjerg Plads 3, 2000 Frederiksberg, Denmark
a r t i c l e
i n f o
Keywords:
Nonnancial performance measures
Target setting
Coordination
Planning
Case studies
Manufacturing
a b s t r a c t
This paper presents an analysis of the resolution of organisational externalities through
the use of nonnancial performance measures for planning. Using a comparative case
study, this paper illustrates how centralised controllers choice of nonnancial performance
measures and target setting in two companies provides critical information to decentralised agents regarding how to balance their performance with the performance of other
decentralised agents in their organisation. This work complements current management
accounting research in that it focuses on the role of nonnancial performance measures
with respect to the design of performance plans for decentralised agents that can be used
to internalise externalities. To date, discussions of externalities in management accounting
research have primarily focused on how performance measurements can be used as a price
mechanism to provide decentralised agents with incentives to internalise externalities. In
addition, this case study illustrates some of the difculties related to acquiring general
knowledge about the externalities of nonnancial performance measures and, therefore,
about whether specic nonnancial performance measures are appropriate for a particular
type of organisation.
2009 Elsevier Ltd. All rights reserved.
1. Introduction
Externalities and their implications have been discussed in management accounting research for decades
(Solomons, 1965; Merchant, 1989; Kaplan and Atkinson,
1998). Various methods to resolve externalities, including
cost allocations (Zimmerman, 1979; Merchant and Shields,
1993; Zimmerman, 2006), the aggregation of performance
measures (Bushman et al., 1995; Keating, 1996), composite performance measures (Dent, 1987; Bouwens and
Van Lent, 2007; Baiman and Baldenius, 2008) and subjective performance measures (Gibbs et al., 2004), have
been observed in practice and discussed in theory. All
are design initiatives used in performance measurement
systems to internalise the external effects that one organisational entity may have on another.
E-mail address: ah.om@cbs.dk.

1044-5005/$ see front matter 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.mar.2009.12.001
Regarding the discussion of nonnancial performance

measures and externalities, the focus in management
accounting research has primarily been on how nonnancial performance measures can be used to measure
externalities, make agents responsible for externalities,
and provide incentives for the internalisation of externalities (Bouwens and Van Lent, 2007; Baiman and Baldenius,
2008). This paper, in contrast, addresses externalities that
nonnancial performance itself causes and analyses how
the value of the measured nonnancial performance is
affected by the externalities that it causes. The paper
aims to illustrate the possibility of resolving externalities through centralised performance planning as well as
to explain how performance targets provide agents with
essential information on how to balance their own performance with the performance of other agents in the
organisation.
An externality is a cost or benet that arises when the
actions of one party affect the utility or production possi-
18
A. Hansen / Management Accounting Research 21 (2010) 1739
bilities of another party (Brickley et al., 2004). Regardless

of whether the context is a market system or a hierarchy, externalities create inefciencies in economic systems.
This is because the decision maker is not fully accountable for all the costs and benets associated with his or her
choices. With respect to performance measurement design,
the notion of an externality is useful because it addresses
the effect that performance along one dimension may have
on the costs and benets of another organisational entity
(employee, team, department, division and so on). These
external costs and benets have a signicant effect on the
value of the measured dimension of performance from a
rm-wide perspective and thereby also on what an appropriate target would be for the performance measure.
Two general approaches to coordination distinguish
between the different ways in which management
accounting systems can internalise externalities, namely,
the price-based system and the planning of quantities system (Milgrom and Roberts, 1992, p. 94). The former relies
on the accounting system to provide the employee with
the incentives to make the right decisions (regulation by
the invisible hand). The latter species and communicates a plan for the employee to make the right decision
(regulation by the visible hand). In management accounting research, the internalisation of externalities has been
discussed through price-based approaches, according to
which aggregated measures, cost allocations, composite
measures, and subjective measures account for external
effects in the performance evaluation of the individual
and hence provide the individual with the incentives to
internalise externalities in his or her decision-making. This
paper, however, illustrates how nonnancial performance
measures also play a coordinative role simply by providing
employees and managers with information about how to
perform to internalise externalities when the measures are
elements in performance planning systems.
The rst part of this paper uses a simple microeconomic model to illustrate the planning problem and to
analyse how a rms optimal level for a given dimension
of nonnancial performance is affected by externalities.
In addition, the ways in which economic relationships
encourage adjustments in performance targets as a way
of internalising positive or negative externalities is considered. Two propositions are derived regarding how
externalities resulting from nonnancial performance
measures can be resolved through nonnancial performance planning. The second part of this paper illustrates
the relevance of the propositions using a comparative case
study. The case study contains externalities at two different companies with respect to the same three nonnancial
performance measures. Both companies develop, produce,
and sell measurement technology and are similar in terms
of strategy and a wide range of other context variables. The
similarities between the two organisations under analysis offer an opportunity to illustrate the contingent nature
of externalities in addition to the relevance of the propositions. In both companies, externalities were resolved by
adjusting performance targets set by central planners. Negative externalities also meant that some dimensions of
nonnancial performance were eliminated from the performance measurement system.
This paper contributes to existing research by analysing

how performance planning by centralised controllers plays
an important role in reducing problems with externalities and myopia. This analysis supplements discussions in
the management accounting literature in which the resolutions of externalities have primarily been portrayed as
a matter of providing employees and managers with the
right incentives. This paper illustrates that sometimes the
provision of information (through performance planning)
may serve as the sole intervention. In this context, performance targets play an essential role and target setting
relies on an analysis of how some dimensions of performance inuences other dimensions of performance that
subsequently affect rm value.
The reminder of the paper is structured as follows. Section 2 contains a conceptualisation of the economics of
nonnancial performance and externalities, which is the
basis for outlining two propositions concerning how the
planning of nonnancial performance can resolve externalities. Section 3 introduces the comparative case study.
Section 4 presents an analysis of the externalities associated with three nonnancial performance measures and
their effect on the design of performance measurement
systems. Section 5 presents a summary and discussion
regarding the case study ndings as well as a reection on the contingent nature of nonnancial performance
externalities and the value of nonnancial performance
measures in organisations. The paper ends with concluding
remarks in Section 6.
2. Nonnancial performance measures,
externalities, and value creation
In this section, the problem of nonnancial performance
externalities is conceptualised, and the use of planning
and target setting as methods for resolving externalities is explored. The section begins by illustrating how
nonnancial performance externalities are created due to
interdependencies among organisational tasks. Second, the
different ways in which management accounting systems
may resolve externalities are briey outlined, with planning as one example. Third, the economics of externalities
are conceptualised and the ways in which optimal levels of
nonnancial performance can be determined with respect
to externalities are discussed. The section ends with an
outline of two propositions that suggest how externalities might be resolved through adjustments in nonnancial
performance targets.
2.1. Externalities and interdependencies among tasks
The notion of externalities is generally used to explain
market failures in welfare economic theory; that is, externalities are positive or negative effects that one economic
agents action have on anothers welfare that are not regulated (Milgrom and Roberts, 1992, p. 75). However, the
notion of externalities may also be used to identify and
characterise inefciencies within a rm. There has been a
long tradition in the management accounting literature of
using this concept to specify the effect of one divisions performance on another divisions performance (Solomons,
Fig. 1. Interdependence between tasks 1 and 2.
1965; Merchant, 1989; Kaplan and Atkinson, 1998). However, this notion may be used at any level of organisation
in rms to specify the costs or benets that any unregulated decisions made by any organisational entity may
have on another entity such as an employee, team, functional department, or division (e.g. Roberts, 2004; Brickley
et al., 2004). In this paper, an externality is dened as the
effect that the performance of one employees, groups, or
divisions organisational task has on the performance of
another employees, groups, or divisions task. From an
internal control point of view, the challenge is to internalise the externality so that the external effect is taken
into account by the affecting party when decisions about
performance are made.
2.1.1. Different types of interdependencies
Consider two tasks, task 1 and task 2. Each is performed
by a different employee, team, group, or division, that is,
A or B. If As performance of task 1 affects Bs production
possibilities for task 2 in a way that is not regulated by the
internal control system, then an externality exists. In this
case, task 1 is the affecting task, and task 2 is the affected
task (see Fig. 1).
Externalities are caused by interdependencies between
tasks or activities.1 Scott (1992) denes interdependence as
the extent to which items or elements upon which work
is performed or the work processes themselves are interrelated so that changes in the state of one element affect the
state of the others (Scott, 1992, p. 230). Interdependencies
imply that the production possibilities of the affected task
or activity may be inuenced in either a positive or negative
way. If the improved performance of one task has a positive
effect on the performance of the other task, the externality
is positive (i.e., a synergy or complement). On the contrary, if the improved performance of one task reduces the
performance of another, the externality is negative (i.e., a
trade-off or substitution).
The interdependence of the tasks giving rise to an
externality may be caused by the fact that the affecting
tasks have an effect on the input, process, or output of
1
Interdependence is a precondition for an externality. However, interdependence among tasks is not necessarily an externality. Thompson,
for instance, distinguishes between three different forms of interdependence, namely, pooled, sequential, and reciprocal (Thompson, 1967).
Pooled interdependence represents a situation in which the same resource
pool performs two tasks. A situation with pooled interdependence cannot
cause an externality problem because an externality is caused when two
resource pools affect each other.
19
the affected task, thereby inuencing production possibilities. These different connection points can be illustrated
through various examples from the literature. For example, Kaplan and Atkinson (1998) describe an input effect
in which the output of one department affects the input
of another department. Suppose the reduced quality of an
intermediate product delivered by department A to department B has a signicant effect on the costs in department
B due to the rework or scrap that the reduced quality
causes. Department A creates negative externalities whenever the quality level of the intermediate product (or
input) is too low for department B (Kaplan and Atkinson,
1998). Zimmerman (2006) provides another example of
how externalities arise when a task in one department
affects the output of another task in another department.
He describes how the hiring of sales personnel in the
sales department affects the output of the Human Resource
(HR) department. Whenever an additional sales person is
hired, the HR department faces additional administrative
work. The HR departments service or output concerning the rest of the organisation temporarily decreases, as
the administrative activities related to the new hire have
the highest priority (Zimmerman, 2006). Another example
illustrating how the affecting task may impact the process
of the affected task can be seen by studying development engineers. In this case, the design of new products
will affect their manufacturability (see for example Ulrich
and Eppinger, 2000). For instance, if development engineers design products with an extraordinarily high number
of components, then the manufacturing processes of the
products are likely to be more complex and have longer
durations. Thus, an organisational entity may create an
external effect on another entity by affecting what the other
entity does (output), how the other entity does it (process),
and with what the other entity does it (input). These relationships are illustrated in Fig. 2 below.
Furthermore, externalities can refer to horizontal as
well as vertical interdependencies between tasks. Thus, an
externality may refer to an interdependency among tasks
at different levels in the hierarchy or tasks linked laterally.
This paper focuses on externalities that relate to horizontal
interdependencies. However, many of the principles discussed here also apply to vertical interdependencies.
2.2. Resolutions of externalities
Externality resolutions have been discussed as long
as externalities have been described. In management
accounting research, one approach has been to include
the external effects in the performance measures used for
evaluating and compensating agents (Kaplan and Atkinson,
1998; Zimmerman, 2006). The idea is to design a performance measurement system so that a negative external
effect created by an agent affects the performance evaluation and compensation of the agent in a negative direction.
Similarly, a positive external effect is designed to have a
positive impact. This implies that an agent pays a price
for creating a negative external effect or receives a reward
for a positive external effect. The agent will internalise the
external effect in his or her decision-making if the agent
seeks to maximise his or her payoff. How much the agent
20
Fig. 2. Different connection points for interrelations between tasks 1 and 2.
will reduce (or increase) the creation of external effects

depends on the price the agent pays (or the reward he or
she receives) for the external effect and how these costs
(or benets) relate to the payoff the agent receives from
performing along the dimension that produces the externalities. Thus, the agent is left to evaluate his or her own
actions. Nevertheless, the punishment or reward of the
externality is the key to aligning the interests of the agent
and owner of the rm and internalising the externality.
Such an indirect way of coordinating actions is often
referred to as a price-based coordination system (Milgrom
and Roberts, 1992, p. 94). In a price-based system, the
rm attempts to guide an agents decisions via price signals, thereby counting on the agent to respond by picking
the appropriate level of performance (coordination by an
invisible hand). An alternative to this form of coordination is the system of quantity planning (Milgrom and
Roberts, 1992, p. 94), whereby the central coordinator simply species the agents performance level (coordination by
a visible hand).
Management accounting research has focused on pricebased resolutions in which performance measurement
system design is used to provide the incentives to
internalise externalities. Quantity planning through performance measurement systems has not received much
attention in the area of internalising externalities. The
following paragraphs contain an outline of these two
approaches.
2.2.1. Internalising externalities by adjusting
performance evaluation and compensation
Externalities can be internalised in the affecting entitys
performance records by the use of a wide range of management accounting techniques (cost allocations, aggregated
performance measures, nonnancial measures and subjective measures). These techniques can be used to punish
or reward agents for their externalities, thereby providing
incentives. The rationale behind the use of each technique
is briey outlined below.
Cost allocation is often referred to as a means of taxing negative externalities. For instance, in reference to
Cooper and Turney (1988), Kaplan and Atkinson (1998)
describe how a single cost driver, such as the number of
parts, might be used to tax complex product design and
encourage engineers to focus their attention on reducing
the number of parts and subsequently prevent negative
effects on purchasing and manufacturing. The number of
parts then becomes a price signal that an agent can take
into account when deciding how many new part numbers
to include in a new product design.2 Another type of cost

allocation that can be used for internalising externalities
is non-insulated cost allocation (e.g. Zimmerman, 2006).
In a non-insulated cost allocation scheme for two divisions, the allocation of corporate-sustaining overhead
costs to one division depends on the operating performance of the other division and vice versa. This provides
incentives to cooperate. Each division takes into account
its effects on the other in its decision making, since the
affecting division is punished by an increased share of
overhead if it reduces the operating performance of the
other division.
Another way to internalise externalities is simply to
increase the aggregation level of the agents performance
measure. For example, division manager performance
could be measured according to rm accounting metrics rather than division accounting metrics (Bushman
et al., 1995; Keating, 1996). Aggregating the agents performance measure implies that the agent will pay for (or
benet from) his or her negative (or positive) externalities to the extent that the external partys performance
is included in the aggregated performance measure.3
Furthermore, a nonnancial performance measure that
captures an externality either by measuring the performance that caused the externality or by measuring the
performance of the external party affected is also a way to
internalise externalities. An example could be the number of parts that development engineers use in their new
product design. In this case, an increase in the number of
parts will negatively affect the development engineers
performance evaluation and compensation. This is not
because costs are allocated to the nancial performance
evaluation of the agent as discussed above but rather
because the additional nonnancial measure is weighted
in the composite measure of the agents performance
(Ridgway, 1956). Thus, the price the agent pays for the
externality is the negative impact on his or her perfor-
2
It is of course a challenge to set the overhead rate (that is, the price
of the externality) correctly. Zimmerman (1979), for example, illustrates
some of the conditions for and effects of the overhead rates underestimation or overestimation of externality costs and to what extent the rm
will benet from an overhead allocation in these cases.
3
However, this incurs other costs: it exposes division managers to
additional compensation risk because the agent becomes dependent on
the performance of other agents as well. Furthermore, the sensitivity of
the performance measure for the agent decreases when the performance
measures aggregation level increases (Banker and Datar, 1989). This is one
of the reasons why adding nonnancial performance measures to the division managers performance measure to create a composite performance
measure has been suggested (Baiman and Baldenius, 2008).
mance evaluation and compensation.4 When it comes to

adding additional measures, another idea is to include
non-insulated performance measures that account for
the external partys affected performance. For instance,
one could include reduced customer satisfaction or sales
for the affected party. Recent empirical and analytical
research has emphasised the role of nonnancial performance measures as a means of providing incentives
for cooperation between agents and mitigating externalities (Bouwens and Van Lent, 2007; Baiman and Baldenius,
2008).5
Finally, the use of subjective measures is yet another
alternative for resolving externality issues; for instance,
upper managers may evaluate the abilities of lower managers to internalise externalities subjectively. Research
has emphasised the role of subjective measures for making measures more complete and able to account for
interdependencies among organisational tasks (Baker et
al., 1994; Gibbs et al., 2004; Moers, 2005).
2.2.2. Internalising externalities through planning
Planning is yet another mechanism that can be used
for internalising externalities that stands in contrast to
the priced-based mechanism described above (Milgrom
and Roberts, 1992). In this case, the central planner simply species the appropriate performance level when
taking externalities into account. The role of nonnancial performance measures in planning is straightforward
because the performance target can be used to communicate the planned level of performance in operational terms.
However, in order for the planner to specify the appropriate level of performance for a nonnancial performance
measure that incurs externalities, the marginal costs and
benets of the nonnancial performance must be known.
The microeconomic relationships that lead to the specication of the appropriate level of performance are briey
outlined in the next section.
2.3. Benets and costs of measured performance and
externalitiesunderstanding value creation related to
nonnancial performance measures
In this paper, the value of a nonnancial performance
measure is conceptualised by the benets received by the
rm from the measured performance and the rms costs
of obtaining it.6 If x is the measured nonnancial performance, c(x) can be dened as the cost of attaining the
measured nonnancial performance for the rm, and b(x)
4
The challenge is, as in the case with cost allocation, to determine the
prices of the externality for the agent and that involves decisions about the
weight of the additional measure in the composite performance measure.
As in the case of cost allocation, underestimation or overestimation of the
price of the externality is also possible in this case.
5
However, earlier studies illustrate the role of composite measures for
cooperation; for instance, Dent (1987) studied a company in which product development managers were held responsible for the sales revenues
of the products they developed and sales managers were held responsible
for the development costs of the products they sold.
6
Value for the rm is presumed to be calculated at any specic time by
discounting to present value the future cash ows that the rm is expected
to generate (Ittner and Larcker, 2001; Merchant, 2006).
21
can be dened as the benet of achieving the measured

nonnancial performance for the rm. The value created at
a particular level of measured nonnancial performance, x,
can then be dened as v(x) = b(x) c(x).
This paper singles out one factor inuencing the value
of the measured nonnancial performance, namely, externalities. The effect that an externality has on the value
is considered a cost. This implies that positive externalities that create benets for the rm will be dened
as negative costs and thus serve as receipts for the
rm.
The cost of an externality can be categorised as a transaction cost (Milgrom and Roberts, 1992; Roberts, 2004).7
Other transaction costs caused by other coordination or
motivational problems, like multi-tasking (Holmstrm and
Milgrom, 1991), risk (Holmstrm, 1979), and manipulation
(Jensen, 2003) issues, also affect the value of individual measures. However, these are not included in the
present discussion. Nevertheless, the principles applied in
analysing the costs of externalities are useful in analysing
the consequences of other transaction costs as well.
To more clearly illustrate the distinctive meaning of the
cost of externalities, another type of costs, namely, production costs, are included in the analysis below. Production
costs are often included in discussions of the value of nonnancial performance measures (Ittner and Larcker, 1998b;
Jensen, 2002; Ittner and Larcker, 2003). Thus, the value of
the performance measure can be expressed by the following equation: v(x) = b(x) cp (x) ce (x), where cp (x) is the
production cost, ce (x) is the cost of externalities caused by
the measured performance, and b(x) denotes the benets of
measured nonnancial performance. v(x) equals the value
created for the rm at the level of measured nonnancial
performance, x.
2.3.1. Production costs
Production costs are the costs of the resources consumed
to produce the measured nonnancial performance, which
include resources consumed to increase customer satisfaction or reduce throughput time such as man hours,
investments in technology, and so on.8 This paper follows
7
Milgrom and Roberts dene transaction costs as the costs of running
the system: the costs of coordinating and of motivating. Thus, under the
hypothesis that organisational structure and design are determined by
minimizing transaction costs, both aspects of the organisation problem
affect the allocation of activity among organisational forms (Milgrom and
Roberts, 1992). They also outline ve kinds of attributes that play important roles in transaction costs analysis, namely, asset specicity, frequency
and duration, uncertainty and complexity, difculty of performance measurement, and connectedness to other transactions (Milgrom and Roberts,
1992, pp. 3132).
8
Milgrom and Roberts also distinguish between production costs and
transaction costs (Milgrom and Roberts, 1992, pp. 3334), but they note
that it is not generally true that the sum of production costs and transaction costs are the total cost of the economic system, as the former depend
only on the technology and the inputs used and the latter depend only
on the way transactions are organised. Production and transaction costs
generally depend both on the organisation and on the technology, which
together sometimes make the conceptual separation between production and transaction costs troublesome. However, for illustrative purposes
with respect to the economics of nonnancial performance, the distinction
in this paper is helpful.
22
earlier work in assuming that the slope of the marginal

cost curve for the production costs is positive and constant. This means that the costs of increasing performance
at higher levels of performance are presumed to be higher
than the cost of increasing performance by the same
amount at lower levels. For instance, it is relatively easier
to increase the levels of customer satisfaction or quality (e.g., by reducing the number of errors in the product
line) at lower levels rather than at higher levels of performance. Furthermore, opportunity costs are likely to
increase the more time an individual employee spends on
a given task rather than on another task included in his or
her job description. Thus, in terms of resource costs as well
as opportunity costs, marginal production costs increase
due to the increasing costs of achieving a higher level of
performance.
2.3.2. Benets of measured nonnancial performance
The value added by the measured performance is not
only determined by the costs but also by the benets related to it. Customer loyalty, employee satisfaction,
productivity, and throughput time are all examples of
dimensions of performance that generate benets for a
rm and benets in terms of either increased income or
reduced costs can be outlined for each of these dimensions. In the following, the marginal benet function is
presumed to be decreasing. This implies that for example, receipts received from increased customer satisfaction
are presumed to decrease proportionally with increases in
customer satisfaction.
The two functions b(x) and cp (x) are depicted in Fig. 3a.
To lay the foundation for an illustration of the economic
effects of externalities, a new preliminary value function,
v(x), is dened as v(x) = b(x) cp (x). This function expresses
the value of measured performance by accounting for
both production costs and the benets of measured performance. The value function is depicted in Fig. 3b and
expresses an insight referenced by several scholars that
it is possible to get too much of a good thing (Jensen,
2002; Ittner and Larcker, 2003). This refers to the fact that
the production costs for achieving a particular performance
level can be so high that they more than offset the benets
of reaching that level. In fact, if the marginal benets are
decreasing and the marginal costs are increasing, which are
the implication of the two functions in Fig. 3a, it is possi-
ble to point out an optimal level of performance, x* , that

maximises the value function, as shown in Fig. 3b.
2.3.3. Adding costs of externalities

After dening the value function to illustrate the value
created at different levels of performance when production costs and benets are taken into account, the next
step is to outline the economic effects of externalities.
As mentioned above, the economic effect of externalities
is conceptualised as a cost. Hence, ce (x) is introduced to
express the costs of externalities as a function of measured performance. These costs are positive for negative
externalities and negative for positive externalities. The
economic consequences of measured nonnancial performance externalities can be depicted graphically, as shown
in Fig. 4a and b. Fig. 4a and b outlines a marginal analysis
of the economic effect of externalities. The marginal cost of
externalities, ce (x), is assumed to be a straight line, where
the slope indicates the increasing marginal costs due to
the increasing effect of the measured nonnancial performance on the external task. Two marginal externality cost
neg
functions are depicted in Fig. 4b. The function c e (x) represents a marginal cost curve for a negative externality, and
pos
c e (x) represents a marginal cost curve for a positive externality. The gure also contains the preliminary marginal
value function, v (x), which is the derivative of v(x) from
Fig. 3b. From Fig. 4a, is it clear that as long as marginal value
creation exceeds the marginal costs of externalities of the
measured performance, the rm benets from increased
performance. However, if the marginal externality cost
exceeds the marginal value created, the rm is better off if
performance is reduced. In the case in which the externality of measured performance is negative and represented
neg
. In the
by c e (x), the optimal level of performance is xneg
case in which the externality of measured performance is
pos
positive and represented by c e (x), the optimal level of
.
performance is given by xpos
Fig. 4a illustrates how the existence of externalities
changes the optimal level of nonnancial performance.
The intersections between the marginal value curve and
the two different marginal externality cost curves indicate
different optima. The costs of externalities affect the optimal level of measured nonnancial performance because
they affect the production possibilities of other organisa-
Fig. 3. (a) Cost and benet functions of measured nonnancial performance. (b) Value function of measured nonnancial performance.
23
Fig. 4. (a) Two marginal cost functions of externalities (positive and negative) and marginal value functions. (b) Two marginal cost functions of negative
externalities with different externality coefcients and the marginal value function.
Fig. 5. (a) Benet and cost functions for measured nonnancial performance including and excluding costs of negative and positive externalities. (b) Value
functions for measured nonnancial performance including and excluding costs of negative and positive externalities.
tional tasks at an increasing rate. Thus, if there are negative

(or positive) externalities, the optimal level for measured
nonnancial performance will be lower (or higher) as
compared with a situation with no externalities due to
increased (or reduced) marginal costs when performance
increases.
Fig. 4b illustrates the consequences of different externality intensities. If the marginal externality cost curve,
c (x), is given by ax and a is the externality coefcient, then
higher values of |a| will indicate stronger externalities. If
a = 0, then no externalities exist. In Fig. 4b, two different
marginal externality cost curves are depicted for measured
nonnancial performance. ce I (x) indicates a marginal cost
curve with a relatively low externality coefcient, while
ce II (x) indicates a relatively high externality coefcient.
It is clear from the cost curves, in Fig. 4b, that the relevant range is reduced for a performance measure with
externality costs represented by ce II (x) compared with a
performance measure with externality costs represented
by ce I (x) where the relevant range is dened as the range
in which measured performance creates value for the
rm.9
9
Zimmerman also uses the notion relevant range, but reserves it for
indicating the rates of output for which the sum of xed and variable
costs closely approximates total cost (Zimmerman, 2006).
These ideas are also reected in Fig. 5a and b, which

depict the total benet, cost, and value functions. Fig. 5a
reproduces Fig. 3a but adds negative and positive externalities to the cost curves. The respective total value
functions for each of these scenarios are depicted in Fig. 5b.
Reviewing the value functions that include the costs of
externalities, it is clear that negative externalities reduce
the value of performance along the dimension measured.
Furthermore, a positive externality may increase the value
created along the measured dimension of performance.
This affects the optimal performance levels and the range
within which performance along the measured dimension
is valuable for the rm.
2.4. Externalities and resolutions due to planning and

target setting
Based on the discussion above, this section proposes an
explanation for how externalities of nonnancial performance might be resolved by quantity planning. The aim is
to suggest how externalities may affect target setting and
the value of the performance measure as well as indicate
potential adjustments to the performance target if externalities are to be internalised.
By adjusting the performance target that matches the
altered optimal performance levels that externalities provide (see Figs. 4a and 5b), central planners can internalise
24
the consequences of externalities. A negative externality

implies that the performance target should be adjusted
downwards relative to a similar situation with no externality (see Figs. 4a and 5b). The question of how much
the performance level should be adjusted to completely
internalise the externality and thus maximise the value created depends on the marginal costs of the externality as
well as the marginal value of performance measured prior
to taking externalities into account. In contrast, a positive
externality implies that the performance target should be
adjusted upwards relative to a similar situation with no
externality. The additional value created along the dimension of performance measured due to the positive effects
on the affected party could be exploited by increasing
the performance level. Overall, this leads to the following
proposition regarding target setting as a means of internalising externalities when performance targets are set by
central planners to reect a rms optimal performance
level:
Proposition 1. Adjusting the performance target downwards (or upwards) from an optimal performance level in
which externalities are not taken into account can be used
as a means of internalising a negative (or positive) externality of a performance measure. This will increase the value
of the measured dimension of performance, ceteris paribus.
The inclusion of ceteris paribus is important because it
species that other effects that the adjusted target might
have, such as effects on employee motivation, are not
included in the present analysis. The only concern here is
the role of performance targets with respect to the internalisation of externalities.
Furthermore, there may be situations in which the
externality coefcient is so high and so many externality costs are incurred that negative externalities almost
totally overwhelm the value of the nonnancial performance measure (see Fig. 4b). When v(x) tends towards zero,
the relevant performance range gradually disappears, and
thus, the company is probably better off without the measure. In addition, measurement costs and other types of
costs related to the measure would probably not be covered by the residual value of the measure in these cases.
Consequently, the following is proposed:
Proposition 2. When the negative externalities that occur
are so high that they eliminate the value of performing
along the measured dimension of the nonnancial performance, the rm will be better off without any performance
along this dimension, and thus, the performance measure
will be eliminated, ceteris paribus.
Again, ceteris paribus is included to acknowledge that
the performance measure may serve purposes other than
coordinating the actions of agents. However, these effects
are not included in the present analysis.
One might also imagine a situation in which positive
externalities (i.e., negative externality costs) are so strong
that the externality coefcient is higher than the preliminary value coefcient of the individual dimension of
measured performance. This would imply that the rm
always benets from an increase in that particular dimension of performance. The negative preliminary value faced
when performing at a high level will be more than offset by

an even higher benet from the positive externality. Thus,
in this case, v(x) tends towards innity. However, because
this is an imaginary situation and that it would never imply
that the measure is removed from the performance measurement system, this paper does not further develop this
case.
3. Introduction to the case study
The purpose of this section is to introduce the case study.
First, the aim of the case study is outlined. Next, the two
companies that are included in the study are briey introduced. Third, the three nonnancial performance measures
under analysis are presented, and the purpose of nonnancial performance measurements in the two companies is
discussed. Finally, the data collection method on which the
case study is based is described, and the principles for data
analysis are outlined.
3.1. The aim of the case study
The aim of the case study is to explore externalities of
nonnancial performance measures in practice and to illustrate the relevance of the propositions outlined above with
respect to resolving externalities through performance
planning. The case study was designed as a comparative
study, as this provides an opportunity to compare the
externalities of the same nonnancial performance measures across two organisations and thereby more clearly
illustrate the factors that cause externalities (and hence the
value) of the performance measures within an individual
organisational setting.
The two companies included in the study were very
similar. Both companies developed, produced, and sold
measurement systems used to measure particle movements in terms of both direction and speed, but the
two companies were not competitors. The rst companys
products were used for measurement of turbulence in air
and gases, while the second companys products were used
for measurements of ows in chemical uids and powder. Nevertheless, the companies products consisted of
many of the same parts, such as probes, sensor modules,
transmission technologies, software modules, and different types of ttings for product set-up. Furthermore, the
companies had very similar competitive strategies and
markets, and the organisation, size, culture, and supplier
relations for the two companies were also alike.
These similarities might lead to the expectation that
the same nonnancial performance measures would t
for both organisations. However, the comparative case
study illustrates the difculties encountered when trying to produce general knowledge about the t (or lack
of t) of nonnancial performance measures in similar
organisations, an objective that is often pursued in the
so-called best practice literature. By focusing on a single
element that affects the organisational t of the individual performance measure (i.e., the performance measures
externalities) this comparative case study revealed how
details related to the product technologies and operations determined interdependencies among jobs in the
organisations and thus the externalities and value of the

performance measure. Thus, the fact that the externalities
of an individual performance measure are hard to generalise implies that the t of a nonnancial performance
measure may also be hard to generalise.
Within the two companies, the implementation of new
nonnancial performance measures was considered an
important element in the execution of their lean manufacturing strategies, which is in line with observations made
in contingency theory studies (e.g. Daniel and Reitsperger,
1991; Chenhall, 1997; Fullerton and McWatters, 2002).
Both companies made signicant changes in their job
designs and production planning according to the principles of lean manufacturing (Womack et al., 1991; Womack
and Jones, 2003) and the implementation of new nonnancial performance measures came as a response to those
changes.
Several new nonnancial measures were proposed in
the two companies to support the move towards lean
manufacturing. Some of these measures were related to
production processes, such as quality, process time, and
productivity measures. Other measures were oriented
towards supplier relations and focused on delivery time
and quality. Still others were focused on customer relationships by measuring delivery time and service. Finally, a
group of measures related to design-for-manufacturability
(DFM) was also considered an important ingredient in realising a lean manufacturing strategy. This paper, however,
only focuses on three nonnancial performance measures,
namely, one manufacturing measure and two designfor-manufacturability measures. All three measures were
debated in both organisations, and the issue of externalities
played a signicant role with respect to all measures.
25
leaders. The two companies, hereafter called Company A

and Company B for anonymity, matched one another not
only in terms of competitive strategy but also in terms of
a wider range of other organisational context variables.
Table 1 below summarises the similarities between the
two organisations. Although the aim of this paper is not
to analyse the characteristics of the two organisations
with respect to the general variables any further, outlining the similarities with respect to the general variables
may be helpful in terms of illustrating the details that
create the externalities in the individual organisational
settings.
3.2.1. Company A
Company A was a world-class manufacturer of instruments that measure ows in air and gases (e.g., turbulence).
The company was relatively small with a turnover of about
62 million Euros, and it operated in a worldwide market
in which it sold to R & D departments in different industries, universities, and other research institutions. There
were three signicant manufacturers in the market, with
Company A holding a dominant position.
The companys competitive strategy was based on a
unique and innovative technological platform on which
the companys products were built. The company was also
skilled at customising its products to specic customer
needs. This skill is a particular focus of this paper. This
company was considered to be a supplier of solutions to
individual customers rather than a manufacturer of traditional products. For instance, the company prole states
that
At [Company A], we believe in providing solutions and
solving problems. Since the company was founded in
1948, thousands of organisations and companies worldwide have depended on the quality and reliability of
[Company A]s products and services to solve their problems (Excerpt from the company prole).
3.2. The two companies

The two companies that were included in this study
developed, produced, and sold measurement technology.
Both were considered high-tech companies and market
Table 1
Characteristics of the two companies.
Company A
Company B
Strategy: differentiation by customisation of high-tech products.

Product: measurement systems for analysing ows and turbulence
in gases and air.
Customers: R&D departments and universities related to a wide
range of industries.
Customisation: adjustments of product parts, programming of
software, and adding additional parts from suppliers.
Environment: market leader, two major competitors, specic
customer measurement problems are considered to be relatively
difcult to predict. Flexibility is considered to be a key to solve
customer measurement problems.
Production system: mounting and assembly, pull production
principles, and high degree of outsourcing.
Organisation: decentralisation of decision rights to workers and
engineers, multi-functional skills among workers in the production
system and team organisation, and cross-functional development
teams.
Culture: high commitment culture; professional culture, but
oriented towards the business and the overall goals of the
organisation.
Turnover: D 62 mill.
Strategy: differentiation by customisation of high-tech products.

Product: measurement instruments for analysing ows in chemical
uids and powder.
Customers: R&D departments in companies and research
institutions related to the chemical and pharmaceutical industry.
Customisation: adjustments of product parts, programming of
software, and adding additional parts from suppliers.
Environment: market leader, three major competitors, specic
customer measurement problems are considered to be relatively
uncertain and difcult to predict. Flexibility is considered to be key
to solve customer measurement problems.
Production system: mounting and assembly, pull production
principles, and high degree of outsourcing.
Organisation: high degree of decentralisation for production
workers and engineers, team organisation and multi-skilled
workers, and cross-functional development teams.
Culture: high commitment culture, professional culture but

oriented towards the business and the overall goals of the
organisation.
Turnover: D 70 mill.
26
That sales engineers had opportunities to adjust, supplement, and program the companys products played a
major role in customising the companys products to the
individual needs of customers, thereby solving individual
measurement problems with respect to customers.
The companys considerable focus on supplementing its
core products with customised solutions made it necessary
to outsource large parts of the companys production processes to subcontractors, as it was not possible to produce
all components and elements necessary for an appropriate
solution for a customer in-house. However, the company
retained its core processes. These comprised the nal production phases of manufactured product units, typically
including assembly, tting, and certain highly specialised
production processes.
The organisation of the manufacturing system in Company A was heavily inuenced by the principles of lean
manufacturing and organisation (Womack et al., 1991;
Womack and Jones, 2003). The ideas of customer orientation, value-stream mapping, empowerment, and pull
production principles were thought to play critical roles in
achieving an integrated manufacturing system. Production
runs were not planned but rather were initiated by kanbans. Employees were given responsibility for productivity
as well as the quality of the product. The plant manager
was in charge of several production teams, and each team
consisted of multi-functional workers that mounted and
assembled the product parts. The plant manager argued
that the multi-functionality of employees made the process ow more quickly than if the product part had to pass
through several functional teams.
Furthermore, decentralisation was used extensively
throughout Company A. Engineers in production, sales,
and the new product development department had significant responsibilities for the development of new products
in cross-functional development teams, and as mentioned
above, production workers were responsible for the dayto-day production scheduling and execution of assembly
and mounting.
In general, employees and managers in Company A were
considered to be highly committed to the overall strategy of
the rm. The integration of customer needs, product development, production, and supplies that is often mobilised
under the lean manufacturing philosophy were assumed
to be a key to the companys success. The production
manager in Company A expressed this presumption as
follows:
I believe that employees in our company are highly
motivatedin production, sales, and development. We
have a professional work environment and people have
interesting jobs. On the other hand, I also believe that
we have been successful in communicating that we are
all in the same boat, which means that integration
between the customers needs, development, manufacturing, and our suppliers is decisive.
With respect to the incentive systems in the company,
the production workers and the engineers in the company
did not have any bonus contracts for individual performance measures. That is, compensation for both workers
and engineers was based on a xed salary; any yearly
bonuses were determined by the companys overall performance. The top managers in the company, however, each
had individual bonus contracts that included prot, sales,
and product development targets.
3.2.2. Company B
Company B was a leading provider of measurement
instruments for ows in chemical uids and powder that
were used for highly specialised process analyses. Its customers were typically R&D departments in companies in
the chemical and pharmaceutical industries as well as universities. Company B had a yearly turnover of 70 million
Euros. It faced intense competition from three other major
competitors in its market, although Company B was considered the market leader. The company developed and
produced high-quality measurement systems, and its competitive strategy resembled that of Company A, as it was
based on an innovative technological platform and extensive use of customisation of products. As with Company A,
the study of Company B focused on its concern for customisation and its aspiration to deliver solutions rather
than products. Company Bs emphasis on customisation
was expressed in the company prole as follows:
Company B provides and supports dedicated and accurate analytical solutions, specically, instruments that
analyse and control specialised pharmaceutical and
chemical processes for the enhancement of our customers business and knowledge creation. (Excerpt
from the company prole)
In Company B, the sales engineers also played a major
role in customising individual products. As with Company
A, the physical adjustment of software programming to
supplement Company Bs product parts was critical in customising the companys product so that the individual
measurement problem with respect to the customer was
solved. As in Company A, customisation was facilitated by
the exible in-house manufacturing system that focused
on core production processes and exible suppliers that
could supply sales engineers with components necessary
for individual customer solutions. The production manager
explained that:
I believe that our manufacturing system is extremely
important when it comes to being exible, and here I
include our internal as well as external manufacturing
system.
Historically, a major part of Company Bs products was
produced internally. However, now only a limited number
of core manufacturing, mounting, and assembly processes
remained in-house. The extensive use of suppliers was
implemented in part to reduce the cost of production but
also to expand exibility as mentioned above, because the
supplier portfolio could supply a wide range of product
parts helpful for sales engineers in customising the companys products.
Company Bs production system was organised according to lean manufacturing principles (Womack et al., 1991).
The production workers were empowered and organised
in production teams with responsibilities for productiv-
27
Fig. 6. Exploring interrelations among three dimensions of nonnancial performance and an external organisational activity.
ity as well as the quality of production. Each worker had

multi-functional skills and was able to handle various different types of machinery in the semi-automated work
environment, which was thought to create job enrichment
as well as make it possible to assign the workers responsibilities for larger parts of the workow. Furthermore,
the development engineers and sales engineers also had a
substantial degree of freedom in their jobs, their responsibilities for new product development, and the continuing
development of the technological platforms in the company. Similar to the idea of freedom under responsibility,
in the words of one engineer, integrating customer needs
with development, production, and supply was considered
critical for the companys competitiveness. Introducing
cross-functional product development teams, kanban production systems, and value-stream analyses were all steps
undertaken to orient workers and engineers towards a
more holistic view of the company.
In Company B, the managers also argued that they had a
highly motivated workforce. Development and sales engineers, as well as people on the shop oor, were considered
to be highly committed. The production manager in Company B stated that:
Overall, I would say that job happiness is high. We have
been good at developing interesting jobs for engineers
as well as for people in production. And people are dedicated to our goal of being a professional organisation
that provides high quality and exible solutions to our
customer.
Finally, the bonus system for the production workers and engineers in Company B was, as in Company A,
based on an overall company performance measure. No
individual performance measures were included in bonus
plans. Neither was there any attempt to include new
nonnancial performance measures in employee bonus
plans.
were an issue for the use of the three measures in both

organisations.10 Furthermore, the external effects of the
three measures were all related to the same type of task
in both organisations, namely, the customisation activities of sales engineer, which made the identication of
externalities more straightforward. Finally, customisation
activities carried out by sales engineers were, as mentioned
above, considered a key strategic concern by both companies, which implies that the externality costs incurred
when this task was affected were signicant and hence
appropriate for an analysis of resolution through nonnancial performance measures.
The three nonnancial performance measures included
in this study include the reduction of components on
printed circuit boards (PCBs) and the reduction of product
parts, which are both performance measures supporting
design-for-manufacturability, as well as probe accuracy,
which is a manufacturing quality measure. The external
tasks that these measures affected are the customisation
activities of sales engineers. Fig. 6 outlines the relationship
between the three measures and the organisational task,
which is the focus of the paper.
3.4. The customisation activities of sales engineers
The sales engineers played a signicant role in both
organisations with respect to customisation. The customisation activities in both companies were carried out by
adding extra components produced either internally or
externally to the major product modules in existence,
programming existing software modules, and/or implementing physical adjustments and modications to extant
product parts. The production manager at Company A
explained as follows:
Our customisation of the products gives us a competitive advantage. We have a sales force that is very good at
understanding the specic measurement problem that
the customer has and then customising the products
3.3. The three nonnancial performance measures and

their use
This paper does not include all nonnancial performance measures debated within and implemented by the
two companies. It focuses on three measures that were
considered in both companies. The reason for including the three measures in this study is that externalities
10
It is worth noting that externalities were not always issues related
to the implementation of the other nonnancial performance measures.
Other factors (e.g., multi-tasking, causality, bias, and so on) brought costs
and benets for other performance measures, but with respect to the three
measures examined in this paper, externalities were the major issue and
hence created a relevant empirical basis for exploring their signicance
for nonnancial performance measurement system design.
28
in ways that match the individual customers specic

needs. . .They do many different things to customise
the products: add new components, adjustments of the
products, software programming.
Similar customisation activities carried out by sales
engineers occurred at Company B. Likewise, Company
Bs production manager emphasised that these activities
played a key role. He explained that:
The sales engineers put a lot of effort in making
our product t the customers individual measurement problem. We supplement our own products with
additional parts, write new software, and adapt the individual product. These are all important activities related
to the process of getting a happy customer. . .. These are
critical issues and I think we are better at doing these
things than our customers.
Thus, there was clear recognition within both organisations of the importance of the customisation activities of
sales engineers.
3.4.1. The choice of performance measures and target
setting as coordination mechanisms
Neither of the companies introduced the new nonnancial performance measures as new elements in employee
incentive contracts, nor did management believe that there
were problems with employee commitment and/or recognition of the importance of integrating different functional
units in the companies. However, managers in both companies experienced issues with respect to the knowledge
among employees regarding what it actually meant to
integrate individual actions with other individuals in the
company. The production manager in Company A had the
following to say regarding this topic:
In general, I think that there is a high commitment
in our companythe employees are motivated. The
challenge for us is to ensure that the employees get
information about what it means to act in the interest of
the company. In our company, a lot of responsibility is
assigned to engineers and workers, and they are willing
to take it, but they do not always know what to do with
it.
Thus, a lack of information rather than a lack of incentives potentially created myopic workers or engineers in
Company A. While managers in both companies recognised that the lean manufacturing principles that had been
implemented had already contributed to integration in
many ways, they also emphasised that nonnancial performance measures played a critical role in informing
employees.
Not only the managers but also the employees recognised the potential of the new performance measures with
respect to information provisions. A production employee
in Company B responded to a question regarding why
he thought that nonnancial performance measurements
would be valuable as follows:
The new performance targets can help us to do things
in the right way. We all want to coordinate, but the
problem is that we are not always sure about what
that means. When performance targets are launched,

it becomes much easier.
A development engineer in Company A characterised
the relevance of nonnancial performance targets and
emphasised the role of controllers with respect to nding
an appropriate level of performance as follows:
Targets are absolutely essential because they inform
us about how to balance our activities with othersfor
instance, our development activities with the sales engineers activities. If I should communicate back and forth
with people to calculate the optimal level of components myself, I would have no time for developing the
new products. It is much better if the controllers do this.
Thus, in the two companies under analysis, controllers
were considered to be the personnel who could set the target and balance the costs and benets of a performance
level in one functional unit with the costs and benets of its
effect on another functional unit. These assessments were
considered the point of departure for setting performance
targets. In other words, there was a call for centralised
information produced by controllers that could inform
decentralised agents how to balance these issues, as these
agents often lacked the time, competences, and/or information available to evaluate the effect of one functional units
interdependence with another on the overall company.
The performance targets were assumed to be equivalent to
informing decentralised agents regarding how much performance was appropriate if externalities were taken into
account.
An alternative to a centralised resolution for internalising externalities might involve negotiations among
the decentralised agents (e.g. Shavell, 2004). However, as
emphasised by the engineer quoted above, a controllers
analysis of costs and benets of a given dimension of nonnancial performance in different functional units from
an overall company perspective was considered a much
more realistic approach than negotiation among decentralised parties; workers and engineers simply did not have
the time and information needed to dene optimal performance levels from an overall company point of view.
The controllers in Company A as well as Company B
were also cognizant of the coordinating role that the performance measures served. The chief controller in Company
A commented on the importance of interdependencies
between organisational tasks when setting targets:
We consider the interdependencies between development, production, and sales when we design the
performance measurement system, and we deal with
the issue when we set the target for the critical success
factors important to us.
Planning was also considered critical in Company B for
dealing with interdependencies among tasks and resolving
issues involving externalities. The Chief Financial Ofcer
(CFO) made the following argument:
I believe we have a powerful tool when it comes to controlling these potential conicts [negative externalities]
if we are careful about the level of performance that we
set in our performance plans and measures.
Thus, managers, engineers, controllers, and workers in

both companies promoted the idea that it was possible to
regulate interdependencies among tasks through the performance measurement system. This fact was the rationale
behind choosing the two companies for further study of
how externalities were internalised through a performance
measurement system.
3.5. Data collection and analysis
The data for the comparative case study included in this
paper originate from two research projects in the electronics industry in Denmark. The rst project involved
carrying out in-depth case studies of management control
and strategy in manufacturing systems in ve electronics
manufacturers, and the second project involved studying
product development, management control, and strategy
in four companies also in the electronics industry. For each
project, data collection was undertaken for approximately
6 months on average, and 1824 interviews were carried out in each company. Furthermore, observations at
meetings and document studies were conducted. Although
each of the two projects initially focused on their own
functional unit, that is, manufacturing and new product development, the interdependencies among suppliers,
new product development, manufacturing, and customer
relationships in the companies necessitated data collection
on a wide range of issues, including management control,
organisational design, and strategy issues in different functional units, in order to better understand manufacturing
and new product development issues in a wider organisational setting.
The relatively broad scope of the studies implied that
eld data on performance measurement and strategy
implementation issues were available for both companies
studied. By chance, two of the companies included in the
research projects were in the process of implementing new
nonnancial performance measures to support manufacturing strategy implementation.
Inspired by the data from the two companies, the
idea for this paper focussing on externalities and nonnancial performance measures in the two companies
was generated. The interviews from the two research
projects provided data relevant to the study of this issue;
however, particularly with respect to Company B, which
was initially part of the product development research
project, some additional information was needed to better explore the research questions in this paper. Thus, it
was decided to carry out a second round of interviews in
the two companies and focus in particular on externality
issues.
In total, 24 interviews were conducted in Company A (18
interviews in the rst round and 6 interviews in the second
round), and 27 interviews in Company B (19 interviews in
the rst round and 8 interviews in the second round). The
interviews from the rst round were conducted at different
hierarchical levels and in different functional units, including purchasing, manufacturing, sales/marketing, product
design, and production engineering. In Company A, the
second round of interviews included the production manager, the sales manager, the chief controller, a production
29
worker, a sales engineer, and a development engineer. The

second round of interviews in Company B included the
production manager, the sales manager, the CFO, the New
Product Development (NPD) manager, a sales engineer, a
production employee, a development engineer, and a business controller.
In both rounds, each interview took between one and a
half to two and a half hours. Each was recorded and transcribed. The interviews were carried out in Danish; the
quotes from the interviews included in this paper have
subsequently been translated into English. With respect to
the observations of the meetings, written summaries and
protocols were produced as documentation.
3.5.1. The role of the propositions
As mentioned above, the interviews from the rst round
generated the inspiration for exploring issues of externalities and nonnancial performance measures. It was
between the rst and second rounds of interviews that
the propositions were generated. The second round of
interviews then provided an opportunity to explore the
propositions in more detail. Thus, the interview guide for
the second round was designed with the propositions in
mind. This made it possible to carefully collect data on how
externalities actually affected the value of the measured
nonnancial performance in the two companies and how
target setting was used as a way to inform workers and
engineers about how to act to balance their performance
with the performance of other units.
The interviews were analysed by exploring respondent
interpretations of the role of nonnancial performance
measures with respect to resolving externalities; naturally, the second round of interviews was more informative
about externality issues than the rst round. Overall, the
respondents agreed in their interpretation of how the measures related to externalities in the companies. It was not
the case that different respondents had different interpretations of the same episodes and issues. Consequently,
it was possible to use the interviews as supplements to
the overall data collection. A pivotal factor for complementarity was the fact that respondents were questioned
about the same measures, externalities, and coordination
issues. The main purpose of the signicant number of
interviews was therefore to obtain more details about the
relationships between the organisational tasks and the
performance measures rather than to explore why different respondents had different interpretations of the
same experience. With respect to externalities, each of
the interviews supplemented one another, and when combined, they created a resource for gaining detailed insights
about the relationship between measures and organisational tasks.
In line with other scholars who have promoted
case study research methods in management accounting research (Scapens, 1990; Keating, 1995; Ahrens and
Chapman, 2006), this paper draws on the idea that case
study research provides an opportunity to study a phenomenon in its real-life setting to illustrate how the
phenomenons signicance in practice is determined by
particularities within the individual organisational practices. The richness of the data on which the case study in
30
this paper is based provides the opportunity to illustrate

how particularities and specicities determine not only
the existence of externalities but also the extent to which
externalities affect the value of nonnancial performance
and how externalities can be managed through target
setting. The case study therefore bridges the rationales produced in both the economic model and the propositions
outlined above in Section 2 with the concrete organisational practice through which managers are confronted
with issues of externalities.
4. Analysis: three dimensions of nonnancial
performance and their external effects on sales
engineer customisation in two organisations
This section reviews the externalities of three dimensions of nonnancial performance that were discussed in
Company A as well as in Company B. The rst dimension
(i.e., probe accuracy) was a quality issue in the manufacturing system. The two other dimensions (that is,
reduction of components on printed circuit boards and
reduction of product parts) were related to design-formanufacturability.
4.1. Probe accuracy as a quality measure
In both companies, quality was a key objective in the
manufacturing system. Total Quality Management (TQM)
initiatives had been implemented, and several quality performance measures had been discussed to implement lean
manufacturing strategies, reduce waste, and increase the
value of company products. One of the measures proposed
and discussed in both organisations that was particularly
interesting for this study was the probe accuracy measure,
which was related to the calibration of the measurement
probes that were essential components in the products produced by both companies. Company A used the probes for
measuring turbulence in uids and gases, and Company
B used the probes for the specialised process analyses in
the pharmaceutical and chemical industries. The quality
of the probe measurements was quantied by using an
accuracy estimate, which was expressed as an estimated
value, plus or minus an uncertainty level. Accuracy was
important because it had a strong impact on product reliability in Company A as well as in Company B. In both
companies, particular jobs were designed as part of the production system to improve probe accuracy. These jobs were
manual and carried out by employees from the production
teams.
When the accuracy of the probes was improved, the
marginal costs related to improving accuracy increased. For
instance, a production engineer in Company A explained:
We can get it almost as accurate as we want. It is just
a matter of how many resources we use. The higher
the accuracy we want, the harder it gets. Getting from
[0.01]11 to [0.001] takes much more than getting from
[0.1] to [0.01].
11
In the following explanation, the scale and level of probe accuracy in
both companies have been indexed and converted to the same perfor-
This was similar to experiences in Company B, in which

increasing marginal costs characterised the improvement
processes as well. As a production worker explained:
The renements that are necessary to get from [o.4]12
to [0.5] are only half of those that I need to get from
[0.04] to [0.05].
When it came to customer decisions to buy company
products, probe accuracy was presumed to be an important parameter in both companies. Customer experiences
with the product and sense of quality were dependent on
accuracy. For example, a sales engineer from Company A
stated:
Going from [0.1] to [0.01] has a remarkable effect on
customers interest in us. If you sell products with a
probe accuracy of [0.1] you only full the needs of a
few customers. Many more customers are interested in
your product if you provide accuracy levels at [0.01].
This quote is also representative of the attitude at Company B. Thus, the benets of accuracy were evident to
everyone in both companies. A low level of accuracy would
lead to few sales orders, whereas high accuracy would lead
to many. However, there was a limit to how accurate the
probes needed to be to affect customer utility. If the accuracy was too high, the customers simply did not recognise
the improvement. A controller in Company B argued that
there was an optimal level of probe accuracy:
If we improve accuracy from, for instance, [0.008] to
[0.006], I do not believe that the customer would recognise it. In addition, it would cost us much extra time
to get there. This is the reason why we have a target
for an accuracy of [0.01]. Beyond this point, it costs us
too much, and the customers really do not recognise the
change. . .It is important that this piece of information
gets communicated throughout the system.
Thus, measuring probe accuracy was considered to be
critical in Company B, and a performance target was set to
ensure that production workers aimed for the optimal level
along this dimension of performance in the production system. The performance target was set at 0.01 in Company
B. The costs of reaching an accuracy level of 0.01 were
offset by the benet of customers being willing to buy company products. On the other hand, performance beyond this
point would be too costly compared to the benets for the
company.
None of the companies experienced any negative externalities related to probe accuracy performance. However,
there was a positive externality related to probe accuracy in
Company A. The implication of that externality was that the
performance target set for probe accuracy in Company A
was higher compared to a similar situation without positive
externalities.
mance level. In practice, the companies use different scales and levels.
However, the simplication made here is unproblematic in terms of
illustrating the reasoning and principles related to setting targets and
internalising positive externalities.
12
Please see the footnote above.
4.1.1. Positive externalities in Company A

The improvement in the accuracy of the probes in
Company A signicantly affected the work processes of
sales engineers related to the customisation of products.
Very high levels of accuracy made customisation easier
because the additional necessary adjustments of Company
As products to address the individual customers measurement problem related to ows in gases or air became
more simple and less time-consuming. The adjustments
were simply harder to obtain and more complex under
lower accuracy levels. As the sales manager of Company A
explained:
The right performance level with respect to the customer is [0.01], but from the sales engineers point of
view, a higher level of performance is desirable. This
is because the adjustments the sales engineers need
to undertake to measure turbulence, for example, are
reduced when the probes accuracy is increased [i.e., tolerance is lowered]. . .. The customer does not really need
accuracy at this level, but we really have an advantage
when we adjust the products, and that is something we
always do.
Thus, improvement beyond the [0.01] level did not
affect the quality of the product from the customer point of
view, but the positive externalities present for the customisation activities of sales engineers eased customisation
processes and thus reduced the total costs for Company
A, because the additional production costs (i.e., additional
time in the manual improvement processes) did not exceed
the cost savings realised by the sales engineers. Tradeoffs were analysed by the controllers in Company A and
communicated through the performance targets. The chief
controller in Company A commented on the economic
rationale for the performance targets as follows:
Of course it is a trade-off. On the one hand, we get
reduced adjustment activities when we increase accuracy. On the other hand, we have the extra resources
needed in production to improve probe accuracy. Probe
accuracy close to [0.001] is the optimal level, we
believe. Beyond that level, it is too costly to improve
it from a production point of view, and the effects with
respect to customisation activities are relatively lower,
which imply that there is no pay-off beyond this level.
Company B did not experience the same positive
externalities as in Company A because the adjustments
necessary to measure the different types of ow issues in
chemical uids and powder that the companys sales engineers were facing were not affected by higher accuracy
levels. In contrast, the adjustments necessary to modify Company As measurement systems were signicantly
reduced at higher accuracy levels.
Thus, in Company A, the positive externality was internalised by increasing the target for probe accuracy relative
to a situation with no externalities. The performance measure was also implemented in Company B. However, in
this case the target was only set by taking customer utility
and production costs into account in the manual production processes. No externalities were involved in target
setting.
31
4.2. Reduction of components on the printed circuit

boards (PCBs)
Yet another set of nonnancial performance dimensions
and their potential customisation externalities were discussed in the two organisations. These were dimensions
related to issues of DFM, which was considered a signicant component of the two companies lean manufacturing
strategies. The rst is the reduction of components on the
PCBs, which is covered in this section, and the second is the
reduction of product parts, which is discussed in the next
section.
In both companies, the number of components on
PCBs in the sensor modules of measurement systems
was considered critical for cost and time consumption in
manufacturing and purchasing processes. The benets of
reducing the number of components in the manufacturing
and purchasing systems were almost the same in the two
companies. In addition, it is generally accepted in the electronics industry that reducing the number of components
on the PCBs reduces the costs in purchasing, inventory,
manufacturing, and so on. A production engineer in Company A stated the following:
It is clear that a reduced number of components is
a benet to us. Machine runs are reduced. Activities
related to inventories and purchasing are simplied.
Furthermore, the production costs of performing along
this dimension (i.e., reducing the number of components)
were relatively easy to trace, as they were related to the
time that development engineers spent on this activity.
As the benets of reducing the number of components
seemed to be relatively higher than the costs, the decision
to include the performance measure in both companies
performance measurement systems was obvious, as it represented a dimension of performance that could improve
value creation for the rm. However, the costs of reducing the number of components on PCBs were not only
related to resources consumed by engineers in the two
companies. In addition, a reduction in the number of
components also affected the customisation activities of
sales engineers. In Company A, the consequences of limiting customisation were so hefty that the controllers
decided to drop the measure. In Company B, the negative externalities were addressed by setting a limit for the
reductions.
4.2.1. Negative externalities in Company A
Engineers in Company A argued that the number of
components on PCBs had serious effects on the customisation processes of development engineers because the
number of components affected the possible measurement
span of the companys products due to the particular transmission technology applied in Company As sensor module.
A high measurement span was pivotal in terms of customising the products because of the uctuation in the air and
gas ows (especially those caused by turbulence) that Company As measurement systems should address. Only one
particular analogue transmission technology was able to
produce a measurement span wide enough to deal with
the uctuations, and the construction of this type of trans-
32
mission technology required a considerably higher number

of components compared with other types of transmission
technologies.
The particular type of technology used in Company
As sensor modules meant that the development engineers had a certain view of the number of components on
PCBs as a performance criterion for new product development. One development engineer made the following
argument:
If a machine can only take 150 different components at
a time, and we decide to use 250 components, then the
machine will just have to run two times.. . . Perhaps it
costs us more, but that doesnt really matter, because it
gives us a unique opportunity to increase the measurement span, which is very important in terms of solving
a customers measurement problem. . .High-scale span
costs us in terms of components! And if we start reducing the number of components, it would cost us in terms
of customisation.
The concern about the particular transmission technology in Company A was also recognised by the chief
controller. According to him, the benets of the reduced
number of components on PCBs, which included reduced
costs in purchasing, stock handling, and manufacturing,
were not as high as the costs of reducing the number of
components, that is, the costs of reducing customisation
opportunities due to the reduced measurement span of the
transmission technology. In fact, it was argued that reducing the number of components on the PCBs would have
quite dramatic consequences on measurement span and
hence customisation, even if only a relatively small number
of components was removed from the design. The production manager explained:
The transmission technology is very sensitive to the
number of components on PCBs. . .Even a relatively limited reduction of components will have a huge impact
on the engineers ability to make products covering
the same measurement span, which again will have
consequences for our customisation. As soon as we
start to reduce the number of components, we reduce
customisation possibilities because we reduce the measurement span.
The negative externalities that a reduction of the number of components on PCBs would have on customisation
activities were considered so strong that the performance
measure was not implemented in Company A. In other
words, the negative externality coefcient for the number of components on PCBs was so high in Company A
that it totally eliminated the potential value along this
dimension. The chief controller in Company A stated the
following:
This dimension is not valuable to us because as soon as
we start to reduce the numbers of components on PCBs,
we pay such a high price in terms of lost customisation.
However, Company B was different in this regard. Performance along this dimension did not have the same
dramatic consequences.
4.2.2. Negative externalities in Company B

In Company B, a reduction of the number of components did not create the same issues as in Company A,
at least with respect to the current numbers of components on the PCBs. The particular type of transmission
technology that was used in Company B in the sensor
module, which was not applicable for Company As products due to the considerable turbulence they needed to
measure, was not as sensitive with respect to the number of components when it came to measurement span.
Furthermore, measurement span did not pose the same
issues for measurements in chemical uids and powder.
Thus, the number of components could be reduced without affecting the customisation processes of sales engineers
signicantly. However, although the measurement span
was not affected signicantly, the reduction could not go
on for too long before the customisation was affected in
another way. The availability of other technical features
attached to the transmission technology that the sales
engineers were using in their customisation activities was
dependent on the number of PCBs in the sensor module.
As the number of components on the PCBs was reduced,
the number of PCBs is also reduced. Initially, an average
of eight PCBs were used in the construction of the sensor module, and it was possible to reduce the number of
PCBs to ve before it affected the availability of the technical features that the sales engineers used to the point
that the benets of the reduction (i.e., manufacturability)
were lower than the costs of the reduction (i.e., lack of customisation). Currently, however, eight PCBs were used in
the sensor module, which implied that there was room for
improvement.
The value creation potential along this dimension
of performance in Company B implied that the performance measure should be implemented there. However,
due to the critical level of ve PCBs, there was a limit
to how many components the development engineers
should eliminate. A business controller explained this as
follows:
The concern for our technical features makes it impossible to use less than ve PCBs when we talk about the
sensor module. If we go below this, the technical features that the sales engineers use will then be reduced
so much that it becomes critical. It means that there is
room for improvement, but also that there is a limit to
how much it is possible to reduce the number of components for customisation.
The controller explained that taking into account the
benets of the reduction of components and PCBs, the costs
related to the resources that the development engineers
consumed on this particular activity, and the potential
affect on the technical features important for customisation, a 15% reduction in the number of components for each
new design of the transmission module was considered
optimal for Company B. The fact that the production costs
of achieving performance are important was emphasised
by the product development manager. If the target was set
above 15%, too much attention would be directed towards
this task relative to the other tasks that development engineers should perform. The development manager stressed
the following:
We do not want the engineers to spend too much time
on it. Therefore, we set up a target of 15% reduction per
new design. I believe that it is a fair target and that it
can be reached within a reasonable amount of time. It
also means that we will not reach the minimum level of
ve PCBs before about three generations from now, and
by that time, there will probably be another technology
available.
Summing up, the transmission technology in Company
B was not affected in the same ways as the one used by
Company A when it came to the reduction of components
on the PCBs. In Company A, the reduction of components
had considerable effects on measurement span, which
again implied that customisation opportunities were dramatically reduced for the sales engineers. In Company B,
the measurement span was not affected by the reduction of
components. However, the availability of technical features
was affected as the number of PBCs was reduced when the
number of components was reduced, and a reduced number of PBCs affected the availability of technical features
in the transmission module. Thus, although the externality costs were not as high in Company B as in Company A,
when it came to the reduction of the number of components on PCBs, it still had an effect on the target chosen for
this particular dimension of nonnancial performance.
4.3. Reduction of product parts

Another performance dimension related to the designfor-manufacturability debated in the two organisations
was the reduction of product parts for each product. The
number of parts used had a large impact on the workload in
the manufacturing and orderprocessing systems of both
companies. A production engineer in Company A described
the issue as follows:
When we develop and design a new measurement system, lots of different parts are included in the system:
several different cables, suspension options, probes,
sensor modules, adaptors, etc. The sales engineers prefer that each different part get its own nished product
code that customers can order separately. It gives them
exibility. However, I oppose this idea, because more
parts mean a greater chance of defects, problems with
stock handling, and larger inventories in the production system. Ideally, we should only have one part per
measurement system. But of course, more parts give us
more exibility to solve the customers particular measurement problem.
In Company A as well as in Company B, a reduction in
product parts was considered benecial because it would
decrease costs in the manufacturing and orderprocessing
systems. Furthermore, relatively few resources were considered for consumption to improve performance along
this dimension, at least in the beginning. The rst reductions could be made by development engineers spending
a couple of extra hours on redesign. However, additional
reductions would soon require relatively more resource
33
consumption in the development department, because it

became relatively more complicated to design part reductions, and additional investments in the product structures
were needed to create new reductions. Thus, the rst set
of part reductions could be achieved with relatively lower
resource consumption per part reduction than the next
set of part reductions, as was argued in both companies.
This provides evidence of an increasing marginal cost curve
when it comes to the costs of implementing part reductions. The costs of part reductions, however, were also
incurred from another side, namely, from the negative
effect that the reduction of parts had on customisation of
the product on the part of sales engineers. This had only
moderate effects in Company A, but in Company B, the
externalities had damaging consequences for the performance dimensions congruence with value creation for the
rm.
4.3.1. Negative externalities in Company A
In Company A, the reduction of parts did affect the customisation activities of sales engineers, but the benets
of reducing the numbers of parts (i.e., reduced costs in
the manufacturing and orderprocessing systems) were
considered to be much higher than the costs of reducing
parts (i.e., increased costs caused by reduced customisation as well as the resources consumed to achieve part
reduction) at least for the rst parts reduced. This implied
that targets for part reduction were implemented in Company A, while the target was considered to be a way
to communicate the balance between the costs and the
benets of parts reduction to the decision makers in the
new product development department. By communicating the target, controllers informed development engineers
about relevant externalities and how to internalise them.
The development engineers appreciated the information
provided by the performance measurement system. A
development engineer commented on the role of the nonnancial performance target as follows:
The number of parts drives manufacturing costs, and I
believe we can improve in this area. . .But, on the other
hand, there is also a limit with respect to the reduction of
parts because of its effect on the sales engineers job. We
can use the target to better understand how to balance
the two concerns.
To a large extent, the target for the reduction of product
parts was determined by the negative effect on the customisation activities of sales engineers. In general, fewer
product parts were considered to imply less customisation.
The limit in terms of how low they could go with respect to
reductions in product parts in Company A was decided by
referring to the main categories of technological features
that the customers should be able to choose from and the
reasonable number of parts that would make it possible to
physically t the measurement system to a customers own
settings without making it too costly for the customer. The
chief controller in Company A explained:
What we did was that we reviewed all the product parts
and the different types of part categories, e.g., sensor
modules, cables, adaptors. We decided on a reasonable
34
number of parts for each category by taking the different

technological situations and physical conditions into
account that our customers were facing. . .Furthermore,
we of course assessed the potential cost-savings in the
manufacturing and orderprocessing systems from the
part reductions and took that into consideration as
well.
Thus, with respect to nding a balance between customisation on the one hand and costs savings in the
manufacturing systems on the other, the controllers in
Company A again became involved in collecting information, analysing it, and communicating the results. The chief
controller described it as follows:
We had an intense dialogue with the sales engineers
on these things as well as with the people in manufacturing. The results of our work are the performance
targets. The targets express how we should balance
these things.
Hence, in Company A, the measure of the reduction of
product parts was implemented because it represented a
particular dimension of performance that created value for
the rm but only up to a certain level. The negative externalities related to performing along this dimension (i.e.,
reducing product parts) had an impact on the extent to
which a certain level of performance paid off for the company. The target was used to communicate the level of
performance at which point further reductions would create more costs than benets for Company A, and this piece
of information was useful for development engineers:
We are very busy and have a lot of things to do. The
targets are useful because they help us remember the
goal related to each part category. I do not remember
these things myself. I have other things to do. I know
that it is a complex job to analyse what the right number
of parts is. Luckily, we have colleagues to do this.
4.3.2. Negative externalities in Company B

In Company B, the reduction of product parts was more
problematic, because the sales engineers experienced a
situation in which further reductions would have considerable negative consequences for their opportunities
to customise the companys products. One sales engineer
commented on the situation:
We cannot reduce the number of parts any more. If
the customers are forced to buy a cable that is too long
simply because we have not designed a cable in a length
that better t this customers needs, we will lose the
order to our competitors. I guess we can say that product
parts really are valuable to us. There are big differences
between the customers systems, and we need a lot of
different parts to deal with the variety.
The controllers agreed on this issue in Company B.
Although they recognised that the number of parts was a
cost driver manufacturing and orderprocessing systems,
they also acknowledged the point voiced by sales engineers that further reductions of the number of parts would
probably be too costly from a customisation point of view
because of the signicant physical diversity across process

systems of customers. A business controller explained the
following:
Well, of course the number of product parts is important in terms of the cost of the manufacturing and
ordering processes.[But] we also need to take the sales
engineers opportunities to create customer value into
consideration. We believe that further reductions in the
number of product parts will have such a negative effect
on the sales engineers customisation that it will erode
the gains made from cost-savings in the manufacturing
system.
The controllers decided to drop the plan of setting a target for the reduction of product parts for the development
engineers, simply because further reductions would have
such a negative effect on the customisation opportunities
of sales engineers. There was no room for improvements,
as one business controller expressed it, thereby taking the
negative effects on customisation into account. The customisation of Company Bs products often required more
hardware adjustments than in Company A, because the
set-up of the products differed more from customer to costumer in Company B due the high physical diversity of the
measurement systems. Thus, engineer adjustment to the
particular systems relied on a large number of physical
components in Company B.
5. Discussion: externalities and nonnancial
performance measurement system design
This section contains a review of the ndings from the
comparative case study of the externalities of the three
dimensions of nonnancial performance. It also contains a
discussion regarding to what extent and under what conditions the propositions outlined in Section 2 were conrmed
with respect to the resolution of the externalities. Furthermore, the roles that nonnancial performance targets
played as coordination devices as well as how information
about externalities was communicated by central planners and controllers in the two manufacturing settings are
discussed. Finally, reections are presented on how the
ndings of this study put the role of strategy into perspective when it comes to choosing nonnancial performance
measures.
5.1. Externalities of nonnancial performance
Table 2 summarises the externalities of the three dimensions of nonnancial performance in the two organisations
and reviews how the measured performance and the external task (i.e., customisation activities by sales engineers)
are related. It also recalls to what extent externalities were
resolved by target setting or by the elimination of performance measures.
The review in Table 2 illustrates that the externalities
of the three dimensions differed from one organisational
setting to another. Even though the two organisations
were alike, the externalities of the performances differed.
The differences were caused by specic technological and
operational elements within the individual organisational
35
Table 2
Externalities of three dimensions of nonnancial performance in two organisations along with their resolutions.
Nonnancial performance
Company
Externalities
Resolution
Manufacturing quality (i.e.,

probe accuracy)
Company A
Positive externalities: extraordinarily high probe

accuracy (i.e., higher accuracy level than what the
customer requests) reduces sales engineer resource
consumption signicantly when adjusting products to
resolve customers individual problems with
measuring ows in air and gases.
No externalities: the customisation activities of sales
engineers are not affected by the probe accuracy level
in any signicant way.
Performance measure implemented

and target setting used to internalise
positive externalities.
Negative externalities: a reduction of the number of

components on PCBs will reduce the measurement
span of the transmission technology appropriate for
measurement of turbulence. A reduced measurement
span will reduce the ability of sales engineers to set up
the companys measurement system to address every
customers ow uctuations.
Negative externalities: reducing the number of
components does not affect the measurement span
because the transmission technology used for chemical
uids is not sensitive to the number of components
applied. However, there is a lower limit for the number
of PCBs (and thus also the number of components) due
to the demand for technical features by customers.
Performance measure excluded; the

negative externalities were too high.
No externalities: the number of product parts only

affects the customisation processes of sales engineers
if the reduction is high. Performance along this
dimension will improve productivity in the
manufacturing and orderprocessing systems without
signicantly reducing customisation up to a certain
level.
Negative externalities: reducing the number of
product parts will limit customisation. The number of
product parts is already low, and the number of
physical product parts has considerable impact on the
opportunities for sales engineers to adapt to the highly
diverse process systems of customers.
Performance measure included and

target setting used to internalise
negative externalities.
Company B
Reduction of components
on PCB
Company A
Company B
Reduction of product parts
Company A
Company B
practices that affected the interdependencies among the

three dimensions of nonnancial performance and the customisation activities of sales engineers (i.e., the external
task).
In Company A, probe accuracy had a positive effect on
the customisation activities of sales engineers, whereas
it had neither a positive or negative effect in Company
B. In this case the difference between the two companies was caused by the adjustments Company As sales
engineers needed to carry out to resolve the customers
individual measurement problem of ows in air and gases.
These adjustment activities were made signicantly easier by higher probe accuracy. Such operational advantages
of higher probe accuracy were not present in Company B,
because the adjustments of Company Bs product to resolve
the measurement problem in chemical uids was not really
affected by probe accuracy. Hence, higher probe accuracy
did not affect customisation in Company B.
The second nonnancial performance dimension, that
is, the reduction of the number of components on PCBs,
caused negative externalities in Company A as well as in
Company B but for different reasons. A reduction of the
number of components on the PCBs would have very negative effects on sales engineer activities in Company A
because the performance of the transmission technology
used in the sensor module in Company A was signicantly
Performance measure implemented.

Externalities do not affect target
setting.
Performance measure included and

target setting used to internalise
negative externalities.
Performance measure excluded; the

negative externalities were too high.
affected by the number of components. Reductions of components was said to reduce the products measurement
span, and a reduced measurement span would seriously
affect the ability of sales engineers to measure all uctuations in air and gas ows (particularly those related
to turbulence) that Company As customers would like
to measure. In Company B, in contrast, the number of
components did not affect the transmission technologys
performance (i.e., measurement span) in the same way
because the type of transmission technology applied in
chemical uids was not as sensitive to the number of components. Nevertheless, another issue came up with respect
to how far the number of components could be reduced
without affecting customisation in Company B. A reduction
of components would affect the number of PCBs, and the
number of PCBs affected how many extra technical features
demanded by customer the sensor module could provide.
Consequently, a minimum level of ve PCBs was set, which
was used as the basis for setting a target for the reduction
in components in Company B.
The third type of performance, namely, the reduction of
product parts, also illustrates how specic circumstances
within the individual organisational setting affect the value
of the measured dimension of nonnancial performance.
In Company B, further reductions in the number of product
parts would have considerable negative effects on customi-
36
sation due to the differences across customer chemical

process systems, which created many diverse physical
measurement system set-ups and thereby a signicant
demand for a high number of product parts. In Company
A, in contrast, the physical set-ups of the measurement
solution in gases and air did not differ to the same extent;
hence, customisation was not that sensitive to the number
of product parts. Furthermore, there was signicant room
for improvement with respect to the number of product
parts in Company A compared to Company B.
A nal point to be made with respect to the creation of
the externalities and interdependencies among tasks is that
the connecting points between the affecting and affected
task differ; see Fig. 2. The reduction of components on PCBs
and the reduction of product parts both affected the input
of the customisation activities of sales engineers (i.e., what
they have to work with), whereas probe accuracy affected
the processes of the sales engineers in Company A (i.e., the
way that they work). This is because extraordinary accuracy meant that sales engineers could eliminate several
steps in their customisation procedures. These different
connection points underline the multiple ways that interdependencies between tasks, and thereby externalities, can
be created in an individual organisational setting.
5.2. Externalities and design of performance
measurement systems
Controllers in both companies adjusted their targets to
internalise externalities. They even decided to withdraw
performance measures with negative external effects that
were too strong. Thus, the case study conrms Propositions
1 and 2 outlined in Section 2 and illustrates how the
propositions outline principles for the coordination of
interdependencies among organisational tasks through
performance measurement and target setting.
5.2.1. Externalities and target setting (Proposition 1)
The comparative case study illustrates how target setting was used in Company A to communicate the synergies
between the high accuracy level of the probes and the customisation activities of sales engineers. It was also used
to protect the customisation activities of sales engineers
against the reduction of product parts in Company B and
the reduction of components on PCBs in Company A by
setting targets that balanced concerns about customisation
with cost-savings in manufacturing and orderprocessing.
Hence, the case study conrms Proposition 1 and illustrates
how target setting was used to internalise positive as well
as negative externalities of multiple dimensions of nonnancial performance.
5.2.2. Externalities and exclusion of performance
measures (Proposition 2)
Externalities also affected the choice of performance
measures in both companies. As suggested in Proposition
2, removing a nonnancial performance measure for which
externality costs are so high that they eliminate the value
of performing along the dimension benets a rm. In one
instance in each company, nonnancial measures were
removed from the performance measurement system due
to very negative effects on customisation. In Company

B, the reduction of product parts had sufciently strong
negative effects on the customisation activities of sales
engineers to the point that it was removed. In Company A,
a reduction of the number of components on PCBs created
such high externality costs that it was withdrawn from the
performance measurement system.
5.3. Nonnancial performance, externalities, and myopic
behaviour: performance measures as planning devices
By denition, externalities are created by myopic managers or employees that disregard the external effects of
their decision-making and thereby make decisions that create less value for the rm than if they had included this
external effect in their decision-making. This paper distinguishes between two types of myopia to further the
discussion of myopia. The rst is caused by the decision
makers lack of incentives to internalise the external effects
in his or her decision-making, and the second is caused by
the decision makers lack of information regarding how the
individuals decisions effect rm value. The myopia of the
latter type was the focus of this comparative case study. In
fact, no incentive problems were considered to be present
with respect to externalities. One employee in Company B
expressed it as follows:
I guess we are always ready to help each other. It
is more a matter of understanding what that actually
means.
Thus, the key role of the nonnancial performance measures in both companies was to provide information that
managers and employees could use to coordinate their
decisions with others.
The fact that it was a lack of information rather than a
lack of incentives that was the issue in the two companies
makes the illustration of the internalisation of externalities somehow simpler. As long as the incentive issue is not
treated as a problem, the decision makers personal costs
and benets of performing along an individual dimension
of nonnancial performance are excluded from the analysis. If the nonnancial performance measures had been
used for the provision of incentives and been included in
incentive contracts, issues related to performance measurement such as manipulability (Jensen, 2003), risk
(Holmstrm, 1979), and completeness (Holmstrm and
Milgrom, 1991) would have inuenced the choice of
performance measures and target setting. However, the
performance measures were decoupled from the incentive
contracts and incentive issues in the two companies. This
implies that it is possible to illustrate the consequences of
a nonnancial performance measures externality directly
as an adjustment of targets and that the externality can be
internalised by informing the individual agent regarding
how much to produce to internalise the externality. Such
a direction or plan is an alternative to incentive contract
or price-based mechanisms (Milgrom and Roberts, 1992)
in which the agent is given an incentive to internalise the
externality by either a gain or loss that the agent receives
from the contract by internalising or not internalising the
externality. The gains and losses related to internalising
or not internalising externalities are price signals that the

agent is expected to act upon and thereby indirectly internalise externalities.
The success of nonnancial performance measures as
coordination devices in the two companies might be caused
by the fact that the myopia problem was only a problem of lack of information and not a problem of lack of
incentive. If the controllers had also needed to resolve an
incentive problem through the performance measures and
had priced each externality issue not only on the basis of
the production of externality costs and benets but also
on costs of manipulability, risk, and completeness, the
situation would have been much more complex and not
necessarily measurable for the controllers.
Thus, this paper illustrates how centralised planning
might also play a role in lean manufacturing, which so
often praises the principle of empowerment (Womack
et al., 1991). The decentralised agents (i.e., workers and
engineers) in the two lean manufacturers included in the
case study requested information about how their own
performance would affect others. These effects were not
something that they had the time or competences to specify themselves. This task was considered to be a centralised
job. However, integration in the two lean manufacturers
was not only achieved through the performance measurement system but target setting and choices of performance
measures played a signicant role in terms of providing
decision makers with specic information about the tradeoffs and synergies between organisational tasks and how
they should be dealt with from an overall rm perspective.
This was apparently not information that the individual
worker or engineer possessed.
5.4. Further reections on the choice of nonnancial
performance measures in general
The central role of an organisations strategy has often
been emphasised when it comes to explaining the choice
of nonnancial performance measures in organisations
(Eccles, 1991; Kaplan and Norton, 1996; Epstein and
Manzoni, 1997; Otley, 1999; Ittner and Larcker, 2003).
Empirical and eld research has illustrated how the choice
of performance measures is linked to the organisations
strategy (e.g. Govindarajan and Gupta, 1985; Simons, 1987;
Ittner et al., 1997; Malina and Selto, 2001; Malina and Selto,
2004), how nonnancial performance measures reect
value drivers of organisations (Ittner and Larcker, 1998a;
Sedatole, 2003; Kaplan and Norton, 2004; Bryant et al.,
2004), and how nonnancial performance measures play a
signicant role in terms of strategy implementation as well
as strategy development (Langeld-Smith, 1997; Bhimani
and Langeld-Smith, 2007).
The results presented in this paper do not contradict
such research but rather add to it as well as putting the
role of strategy into perspective when it comes to choosing
valuable nonnancial performance measures in organisations. By highlighting how the transaction costs (in this
case, the costs of externalities) of performing along a particular dimension of nonnancial performance affects its
value, more issues are added to the analysis of the value
of nonnancial performance measures in the individual
37
organisational setting. Strategy is a factor that affects the

value of the individual dimension of nonnancial performance, not least when it comes to understanding its
benets. However, an analysis of the transaction costs of
performing along the particular dimension of performance
that the performance measure represents should also be
included. This paper demonstrates this by focusing on how
task interdependency among the measured nonnancial
performance and other organisational tasks affects rm
value creation.
This type of transaction cost-based analysis is of course
limited with respect to explaining performance measurement system change in general. Broader frameworks
are often necessary to understand choices and changes
related to performance measurement systems in practice
(Mouritsen, 1999; Vaivio, 1999a,b; Wouters and Wilderom,
2008) in part because interests (Chua, 1995; Briers and
Chua, 2001) and institutional concerns (Granlund and
Lukka, 1998; Burns and Scapens, 2000) are drivers of
accounting change and resistance. Nevertheless, the relatively limited analysis in this paper of the value of three
nonnancial performance measures in the two organisational settings has made it possible to study and illustrate
externalities as the determining factor when it comes to
performance measurement system design. This is part of
the explanation of what drives changes to management
accounting practices in organisations.
6. Conclusions
This paper highlights that the problem of externalities is
not only caused by myopic decision makers with no incentive to internalise the external effect of their decisions. The
problem is also caused by decision makers with the best
intensions that nevertheless lack information about how
their actions affect others or how to correct them to act in
the interest of the organisation. This paper illustrates how
controllers choice of performance measures and target setting plays a central role in terms of providing decentralised
agents (i.e., workers and engineers) with information about
how to act to internalise their externalities.
Furthermore, by means of a case study, this paper
illustrates that despite the coordination mechanisms
present in the two lean manufacturers studied, including
multi-functional skilled workers, cross-functional teams,
value-stream mapping, and pull production, there was
still a need for more information about how to balance
interdepartmental decisions. The study illustrates how
nonnancial performance measures played a critical role
with respect to fullling this need. By emphasising the
nonnancial performance measurement systems role as
a centralised planning system, the paper supplements
the analyses of externality resolutions in management
accounting research, which have been primarily focused
on the design of incentive systems rather than planning
systems.
The planning of nonnancial performance in the two
organisations was based on the principle that if nonnancial performance created a negative (or positive)
externality, the externality could be internalised by adjusting the performance target in a downwards (or upwards)
38
direction. Furthermore, the choice of performance measures was also used for the provision of information
about externalities. When the negative externalities that
occurred from a nonnancial performance measure were
so high that they eliminated the value of performing
along this dimension of nonnancial performance, the
performance measure was not implemented because the
companies were better off without enhanced performance
along this dimension.
An additional point derived from the case study is
that although the two organisations were very similar,
the same measures had different effects in each of the
organisations. This illustrates some of the obstacles that
are encountered when trying to produce general knowledge about which nonnancial performance measures t
best into different types of organisations. The case study
illustrates how the details related to the operational, organisational, and technological sides that often escape general
characterisations of the organisations nevertheless determined the externalities of the performance measures and
thereby their value and t in the individual organisation. The capability to include these details in the analysis
illustrates again how case studies supplement and add
to other types of research methods. The importance of
studying accounting in the context in which it operates (Hopwood, 1983) is emphasised, especially when it
comes to producing specic knowledge of how the externalities are created in practice and how the economics
of multiple dimensions of nonnancial performance are
determined.
Acknowledgements
The author appreciates helpful comments and suggestions from Robert Scapens (editor-in-chief), Ivar Friis,
Jan Mouritsen, Tams Vmosi, two anonymous reviewers
and participants at the First Workshop on Management
Accounting as Organizational and Social Practice (MASOP),
HEC, Paris, 2008.
References
Ahrens, T., Chapman, C.S., 2006. Doing qualitative eld research in
management accounting: positioning data to contribute to theory.
Accounting, Organizations and Society 31 (8), 819841.
Baiman, S., Baldenius, T., 2008. Non-nancial performance measures as coordination devices. SSRN eLibrary available from:
http://ssrn.com/paper=1155562.
Baker, G., Gibbons, R., Murphy, K.J., 1994. Subjective performance
measures in optimal incentive contracts. The Quarterly Journal of Economics 109 (4), 11251156.
Banker, R.D., Datar, S.M., 1989. Sensitivity, precision, and linear aggregation of signals for performance evaluation. Journal of Accounting
Research 27 (1), 2139.
Bhimani, A., Langeld-Smith, K., 2007. Structure, formality and the
importance of nancial and non-nancial information in strategy development and implementation. Management Accounting
Research 18 (1), 331.
Bouwens, J., Van Lent, L., 2007. Assessing the performance of business unit
managers. Journal of Accounting Research 45 (4), 667697.
Brickley, J.A., Smith, C.W., Zimmerman, J.L., 2004. Managerial Economics
and Organizational Architecture. McGraw-Hill, Boston.
Briers, M., Chua, W.F., 2001. The role of actor-networks and boundary
objects in management accounting change: a eld study of an implementation of activity-based costing. Accounting, Organizations and
Society 26 (3), 237269.
Bryant, L., Jones, D.A., Widener, S.K., 2004. Managing value creation within
the rm: an examination of multiple performance measures. Journal
of Management Accounting Research 16, 107131.
Burns, J., Scapens, R., 2000. Conceptualization management accounting change: an institutional framework. Management Accounting
Research 11 (1), 325.
Bushman, R.M., Indjejikian, R.J., Smith, A., 1995. Aggregate performnace measures in business unit manager compensation: the role
of intrarm interdependencies. Journal of Accounting Research 33
(Suppl.), 101128.
Chenhall, R.H., 1997. Reliance on manufacturing performance measures,
total quality management and organizational performance. Management Accounting Research 8 (2), 187206.
Chua, W.F., 1995. Experts, networks and inscriptions in the fabrication
of accounting images: a story of the representation of three public hospitals. Accounting, Organizations and Society 20 (2/3), 111
145.
Cooper, R., Turney, P.B.B., 1988. Tektronix: the portable instrument division. Harvard Business School Cases.
Daniel, S.J., Reitsperger, W.D., 1991. Linking quality strategy with management control systems: empirical evidence from japanese industry.
Dent, J.F., 1987. Tension in the design of formal control systems: a
eld study in a computer company. In: Bruns, W.J., Kaplan, R.S.
(Eds.), Accounting and ManagementField Study Perspectives. Harvard Business School Press, Boston.
Eccles, R.G., 1991. The performance measurement manifesto. Harvard
Business Review 69 (1), 131137.
Epstein, M.J., Manzoni, J.-F., 1997. The balanced scorecard and tableau de
bord: translating strategy into action. Management Accounting 79 (2),
2836.
Fullerton, R.R., McWatters, C.S., 2002. The role of performance measures
and incentive systems in relation to the degree of JIT implementation.
Gibbs, M., Merchant, K.A., Van der Stede, W., Vargus, M.E., 2004. Determinants and effects of subjectivity in incentives. The Accounting Review
79 (2), 409436.
Govindarajan, V., Gupta, A.K., 1985. Linking control systems to business
unit strategy: impact on performance. Accounting, Organizations and
Society 10 (1), 5166.
Granlund, M., Lukka, K., 1998. Its a small world of management accounting
practices. Journal of Management Accounting Research 10, 153179.
Holmstrm, B., 1979. Moral hazard and observability. The Bell Journal of
Economics 10 (1), 7491.
Holmstrm, B., Milgrom, P., 1991. Multitask principal-agent analyses:
incentive contracts, asset ownership, and job design. Journal of Law,
Economics and Organization 7, 2452 (special issue).
Hopwood, A.G., 1983. On trying to study accounting in the contexts in
which it operates. Accounting, Organizations and Society 8 (2/3),
287306.
Ittner, C.D., Larcker, D.F., 1998a. Are nonnancial measures leading indicators of nancial performance? An analysis of customer satisfaction.
Journal of Accounting Research 36, 135.
Ittner, C.D., Larcker, D.F., 1998b. Innovations in performance measurement: trends and research implications. Journal of Management
Accounting Research 10, 205238.
Ittner, C.D., Larcker, D.F., 2001. Assessing empirical research in managerial accounting: a value-based management perspective. Journal for
Accounting and Economics 32, 349410.
Ittner, C.D., Larcker, D.F., 2003. Coming up short on nonnancial performance measurement. Harvard Business Review 82 (2), 8895.
Ittner, C.D., Larcker, D.F., Rajan, M.V., 1997. The choice of performance
measures in annual bonus contracts. The Accounting Review 72 (2),
231255.
Jensen, M.C., 2002. Value maximization, stakeholder theory, and the corporate objective function. Business Ethnics Quarterly 12 (2), 235256.
Jensen, M.C., 2003. Paying people to lie: the truth about the budgeting
proces. European Financial Management 9 (3), 379406.
Kaplan, R.S., Atkinson, A.A., 1998. Advanced Management Accounting.
Prentice Hall, Upper Saddle River.
Kaplan, R.S., Norton, D.P., 1996. Using the balanced scorecard as a strategic
management system. Harvard Business Review (JanuaryFebruary),
7585.
Kaplan, R.S., Norton, D.P., 2004. Measuring the strategic readiness of intangible assets. Harvard Business Review (February), 5263.
Keating, A.S., 1996. Determinants of divisional performance evaluation
practices. Journal of Accounting and Economics 24, 243273.
Keating, P.J., 1995. A framework for classifying and evaluating the theoretical contributions of case research in management accounting. Journal
of Management Accounting Research 7 (Fall), 6686.

Langeld-Smith, K., 1997. Management control systems and strategy:
a critical review. Accounting, Organizations and Society 22 (2),
207232.
Malina, M.A., Selto, F.H., 2001. Communicating and controlling strategy:
an empirical study of the effectiveness of the balanced scorecard. Journal of Management Accounting Research 13, 4790.
Malina, M.A., Selto, F.H., 2004. Choice and change of measures in performance measurement models. Management Accounting Research 15
(4), 441469.
Merchant, K.A., 1989. Rewarding ResultsMotivating Prot Center Managers. Harvard Business School Press, Boston.
Merchant, K.A., 2006. Measuring general managers performances.
Accounting, Auditing and Accountability Journal 19 (6), 893917.
Merchant, K.A., Shields, M.D., 1993. When and why to measure costs less
accurately to improve decision making. Accounting Horizons 7 (2),
7681.
Milgrom, P., Roberts, J., 1992. Economics, Organization and Management.
Prentice-Hall, Upper Saddle River, New Jersey.
Moers, F., 2005. Discretion and bias in performance evaluation: the impact
of diversity and subjectivity. Accounting, Organizations and Society
30, 6780.
Mouritsen, J., 1999. The exible rm: strategies for a subcontractors management control. Accounting, Organizations and Society 24 (1), 3155.
Otley, D., 1999. Performance management: a framework for management control system research. Management Accounting Research 10,
363382.
Ridgway, V.F., 1956. Dysfunctional consequences of performance measurements. Administrative Science Quarterly 1 (2), 240247.
Roberts, J., 2004. The Modern FirmOrganizational Design for Performance and Growth. Oxford University Press, Oxford.
Scapens, R.W., 1990. Researching management accounting practice:
the role of case study methods. British Accounting Review 22 (3),
259281.
39
Scott, R.W., 1992. OrganizationsRational, Natural, and Open Systems.

Prentice-Hall, Englewood Cliffs, N.J.
Sedatole, K.L., 2003. The effect of measurement alternatives on a nonnancial quality measures forward-looking properties. The Accounting
Review 78 (2), 555580.
Shavell, S., 2004. Foundation of Economic Analysis of Law. The Belknap
Press of Harvard University Press, Cambridge, Massachusetts.
Simons, R., 1987. Accounting control systems and business strategy:
an empirical analysis. Accounting, Organizations and Society 12 (4),
357374.
Solomons, D., 1965. Divisional Performance: Measurement and Control.
Financial Executive Research Foundation, New York.
Thompson, J.D., 1967. Organizations in ActionSocial Science Bases of
Administrative Theory. McGraw-Hill, New York.
Ulrich, K., Eppinger, S.D., 2000. Product Design and Development.
McGraw-Hill, Boston.
Vaivio, J., 1999a. Examining the quantied customer. Accounting, Organizations and Society 24 (8), 689715.
Vaivio, J., 1999b. Exploring a non-nancial management accounting
change. Management Accounting Research 10, 409437.
Womack, J., Jones, D.T., Roos, D., 1991. The Machine That Changed the
World. HarperCollins, New York.
Womack, J.P., Jones, D.T., 2003. Lean ThinkingBanish Waste and Create
Wealth in Your Corporation. Free Press, New York.
Wouters, M., Wilderom, C., 2008. Developing performance measurement
systems as enabling formalization: a longitudinal eld study of a
logistics department. Accounting, Organizations and Society 33 (45),
488516.
Zimmerman, J.L., 1979. The cost and benets of cost allocations. The
Accounting Review 54 (3), 504521.
Zimmerman, J.L., 2006. Accounting for Decision Making and Control.
McGraw-Hill Irwin, New York.

Nonfinancial Performance Measures, Externalities and Target Setting A Comparative Case Study of Resolutions Through Planning Hansen2010

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Nonfinancial Performance Measures, Externalities and Target Setting A Comparative Case Study of Resolutions Through Planning Hansen2010

Uploaded by

Copyright:

Available Formats

Management Accounting Research 21 (2010) 1739

Contents lists available at ScienceDirect

Management Accounting Research

Nonnancial performance measures, externalities and target setting:

E-mail address: ah.om@cbs.dk.

Regarding the discussion of nonnancial performance

A. Hansen / Management Accounting Research 21 (2010) 1739

bilities of another party (Brickley et al., 2004). Regardless

This paper contributes to existing research by analysing

A. Hansen / Management Accounting Research 21 (2010) 1739

Fig. 1. Interdependence between tasks 1 and 2.

A. Hansen / Management Accounting Research 21 (2010) 1739

Fig. 2. Different connection points for interrelations between tasks 1 and 2.

will reduce (or increase) the creation of external effects

to include in a new product design.2 Another type of cost

A. Hansen / Management Accounting Research 21 (2010) 1739

mance evaluation and compensation.4 When it comes to

can be dened as the benet of achieving the measured

A. Hansen / Management Accounting Research 21 (2010) 1739

earlier work in assuming that the slope of the marginal

ble to point out an optimal level of performance, x* , that

2.3.3. Adding costs of externalities

A. Hansen / Management Accounting Research 21 (2010) 1739

tional tasks at an increasing rate. Thus, if there are negative

These ideas are also reected in Fig. 5a and b, which

2.4. Externalities and resolutions due to planning and

A. Hansen / Management Accounting Research 21 (2010) 1739

the consequences of externalities. A negative externality

when performing at a high level will be more than offset by

A. Hansen / Management Accounting Research 21 (2010) 1739

organisations and thus the externalities and value of the

leaders. The two companies, hereafter called Company A

3.2. The two companies

Strategy: differentiation by customisation of high-tech products.

Strategy: differentiation by customisation of high-tech products.

Culture: high commitment culture, professional culture but

A. Hansen / Management Accounting Research 21 (2010) 1739

A. Hansen / Management Accounting Research 21 (2010) 1739

ity as well as the quality of production. Each worker had

were an issue for the use of the three measures in both

3.3. The three nonnancial performance measures and

A. Hansen / Management Accounting Research 21 (2010) 1739

in ways that match the individual customers specic

that means. When performance targets are launched,

A. Hansen / Management Accounting Research 21 (2010) 1739

Thus, managers, engineers, controllers, and workers in

worker, a sales engineer, and a development engineer. The

A. Hansen / Management Accounting Research 21 (2010) 1739

this paper is based provides the opportunity to illustrate

This was similar to experiences in Company B, in which

A. Hansen / Management Accounting Research 21 (2010) 1739

4.1.1. Positive externalities in Company A

4.2. Reduction of components on the printed circuit

A. Hansen / Management Accounting Research 21 (2010) 1739

mission technology required a considerably higher number

4.2.2. Negative externalities in Company B

A. Hansen / Management Accounting Research 21 (2010) 1739

4.3. Reduction of product parts

consumption in the development department, because it

A. Hansen / Management Accounting Research 21 (2010) 1739

number of parts for each category by taking the different

4.3.2. Negative externalities in Company B

because of the signicant physical diversity across process